Seeking Info on bow front divided light door

[Peter Quinn]

I may have a job upcoming requiring bow front divided light french entry door pairs, maybe panels and glass too. Anyone ever run across any good links to methods of work for this type of thing? I have a few old theory books but can't find much in the way of real practical applications using modern machines. I'm a researcher by nature, I have some experience with many aspects of this type of work by would love to see somebody that has put the whole method together. Feel free to pm me if a link can't be posted publicly. I have searched several other forums an am coming up with very limited info save one good blog on radius front cabinets, nothing regarding radius front sash or entries. Thanks.

[Mel Fulks]

The old ones I have seen are always veneer over segments for rails . Never seen a bowed French door, the first thing I would find out us if the GLASS has to be bowed. I saw a bowed sash made with sticking circular but rabbet was cut straight for standard flat glass.

[Kevin Jenness]

Complicated prospect for sure. I haven't done such a thing, though I have made some bow front frame and panel cabinet doors. Essentially the same setup as flat with a concave side up cradle to hold the rails for tenoning/coping. If the radius is small enough you might get away with flat stiles. On 18" radius 2 1/4" wide stiles I milled them flat, ran them with a bevel on the moulded edge and a small shim under the side away from the spindle, then blended the curve in with hand tools. The rails were bent laminations. The stiles I suppose could be also, or moulded with convex and concave cutters.

As far as references, the best one I have is Peter Collings' "Curved Work in Carpentry and Joinery", but it doesn't have much to say about machine work. I would suggest you post your question at Woodweb's Architectural Woodworking forum. Perhaps one of the old salts there would respond.

As Mel says, curved or straight glass will make a considerable difference in setups. I daresay it would be easier to mill curved rabbets than straight on curved sash bars, but harder and more expensive to get the glass made accurately, especially if insulated units.

Good luck with it. Be sure to charge plenty.

[Jamie Buxton]

Curved glass (it is called bent glass in the glass world) will be a big issue. It is possible to get glass custom-bent to a specific radius, but it is quite expensive. AFAIK, custom-bent insulated glass doesn't happen.

[david brum]

Peter, if you get the job, I sincerely hope you'll take lots of photos and share them. It sounds like a challenge, for sure. I'd love to see how you work out the joinery for the mullions.

[Peter Quinn]

I'll have to inquire about the glass, interesting consideration, I'm just the mechanic on this one (potentially), I could jig up to set flat panes in a curved frame, that will make the glass stop more interesting to fit! I've only had a brief initial conversation about the work, having never done it I'd like to be as mentally prepared as possible should it happen. I've heard of people using flat stiles on a bowed door and just making them fit, but this is an exterior door in a cold climate with weather seal requirements, or at least I imagine thats a requirement. Having flat stiles would present problems where door meets rabbited jamb head and sill even if the slightly segmented aesthetic were acceptable. I could see having knives made for a Hussey that match the average radius of the whole set to get pretty close, or CNC outsourcing the blanks, possibly machining stave core blanks and veneering those to minimize sanding? Or making a router jig like a giant guitar knock radius jig, one concave and one convex?

I have the Collings book, lots of great methods for doing layout, particularly on curve on curve work, but very little on the nuts and bolts of how to do anything. I poked around at wood web hoping somebody had tackled this prior, and I'll guess somebody there has, but no discussion of it. Maybe time for an inquiry.

[Peter Quinn]

Peter, if you get the job, I sincerely hope you'll take lots of photos and share them. It sounds like a challenge, for sure. I'd love to see how you work out the joinery for the mullions.

That part I'm pretty clear on, the verticals will be flat, no need to radius those, I'm told the top is straight, not arched thankfully, the horizontal bars can be done on a sliding shaper with a cope set and a carriage clamped to the sliding table to hold the work at the correct angle, run maybe a 1/2" stub tenon, very much like traditional sash bars in that sense. Whats baffling me is how to take a traditional straight threshold section and turn it into a radius. Maybe on a hussey, with glue ups, or a plunge router and trammel then lots of belt sanding? I'm seeing that taking more hours than I care to think about. The other tricky parts is making all this bowed horizontal bars, profiling them, then taking out the rabbits accurately, and making the stops. I assume the convex side faces out, haven't seen plans yet. I'm guessing its a few months out, if I get the work I'll definitely try to document construction.

[Jamie Buxton]

Remember: curves cost money. You have a whole shop full of machines that make lumber flat and square, but making lumber curved means templates and special setups and such. It can be done, but it just costs a bunch more time than you might think.

[Sam Murdoch]

I have used this company several times for bent glass work - http://www.standardbent.com
If you give them an accurate drawing of the curve or all the dimensions and the correct radius they will give you good product.

My inclination would certainly be to build to the curve. Depending on the radius (and if you own a bandsaw) you might be able to take the rails from thick stock without the need to laminate. I prefer this because it eliminates any spring back issues. Easy enough to make the horizontal muntins from sawn stock too. Depending on the profile you will need to do some fancy shaper/jig/bearing work but that goes without saying for this type of project.

[Larry Edgerton]

I made a pair to replace a failed attempt by another company years ago. I made the styles out of solid wood. I ground a pair of convex, and a pair of concave knives to fit my Porter Jointer that had tables that could be slid back to allow profiled knives. I used a powerfeed on a custom sled that rode outside the knives to shape the two sides of the styles. The rails I used 1/8" laminations glued together on a form using West System. I used West here because that is where the other doors failed, radius changed over time I suppose due to glue creep, but I am just guessing on that. There was no springback with this ply size on a 2" thick door. These had full bent glass so no mullions so I did not have to deal with that. They are still there and it has been twenty years. Hope this helps.

I believe these were a 12' radius just for reference. Tighter would have required a different strategy.

Larry

[Peter Quinn]

Sam, thanks for that glass link, I'll pass that up the chain if it becomes my work.

Larry, mind my asking how did you handle the threshold? We usually do an adjustable gasket at the bottom of the door, a flat the thickness of the door starts the sill, then a 5/16" vertical drop for water dam, then a 5-10 degree pitch drop to the exterior depending on exposure and wall thickness. But I have no idea how to gasket the bottom on a curve, the stuff I've used only does straights, and actually making a wooden sill with typical profile in curve has challenges, big curved jig on a router trammel and lots of passes? Followed by a jack plane? I'm left scratching head and butt at the same time on that one. If this happens I'm pushing for west systems on this. Did you thicken it or just wet it out as it comes out of the pumps?

[Larry Edgerton]

Brushseals from Conservation Technologies located out in your neighborhood. I use their tubeseals on most doors. Google Conservation Technologies and weatherstrip and it will pop right up.

I made the sill and and head jamb with a 3 1/4hp router on a heavy trammel. I custom ground a HS steel bit for the rain ledge and put a hook in the door bottom so that the brush seal sealed vertically and a tube seal in the door bottom sealed against the threshold. These were inswing french doors so I could do the double seal to combat the winds off of the lake.

I used the West system straight, just like I like my scotch.........

[Peter Quinn]

First third is out the door, so much time pressure I didn't take a single photo...not even of the finished product! Spent probably 120 man hours making jigs. Lots and lots and lots of jigs and fixtures. Turns out it is 9 french pairs, curved glass, SDL grills (which have proven to be more difficult than TDL IMO), transoms. I'm in the middle of the second phase, took a few pics so I can remember the process. Scary stuff!

[Mel Fulks]

Agree on that grill thing, with true divided there is a way to hold it together. We will look forward to more photos.

[Bill Orbine]

I'm already envision buying a few auto/truck windshields.... that is if you could find the proper curve. Maybe these windshield replacement companies will have a bunch cheep. It's a thought.

[Peter Quinn]

No better way to fare a curve than with a compass plane! Card scraper brings up the rear, minimal sanding. Try as I might there are minor thickness variations between stile and bowed rails, running everything face down puts the errors to the concave side where they are tough to sand, the old 113 make it all level quick. CNC....we don't need no stinking CNC! Next step.....muntins. Note the really fancy coping jig in the last picture...took hours to make...cost millions to design.:rolleyes:

[Tom M King]

Glad to see the 113 still getting called on. The only thing remotely close to that that I've done is replace some rotten bottom rails on curved double hungs in a turret on a late 1800s house. It was in the '80s, and before I owned a 113. I ended up Westing some thin strips of White Oak together (sash were painted inside and out), and formed the rabbet for the glass in the layup. The scariest part was handling the old large sheets of curtain glass (I'm guessing) for fear that I would break one. The owners that I did the work for passed on some years back, but every once in a while I drive by that house, and the windows are still looking good, so I guess it's a good testimony for West Systems.

Impressive project!

[Tom M King]

Peter, I just thought about the use of the 113, along with the other stuff, qualifies you to be called a "hybrid woodworker".:rolleyes:

[Peter Quinn]

A few more pics.....a coping sled that doubles as a ski jump or car ramp......curved bars equals curved sleds! This one handles all the curves from the little grill bars to the larger bottom rails. Also some pics of the grills being fitted and processed. Next stop...sticking the horizontals. It involves another curved fence that holds the 7 3/4" bars captive on their journey past the 4" cutter. The odds are not in their favor, but with a good feeder setup and the proper fence it works. Stay tuned.

[John A langley]

This is been a very interesting thread, The doors look great thanks for taking time to post The pictures , would like to see pictures of the shop thanks again

[Tom M King]

My friends that work for NASA say that their definition of an expert is someone who has done something once before. This is, of course, because they are always figuring out how to do something that hasn't been done before. I don't care if you've ever done this before or not. You are absolutely qualified as an expert at building curved doors, and sash.

[Peter Quinn]

My friends that work for NASA say that their definition of an expert is someone who has done something once before. This is, of course, because they are always figuring out how to do something that hasn't been done before. I don't care if you've ever done this before or not. You are absolutely qualified as an expert at building curved doors, and sash.

Im pretty sure I wasn't an expert when I started this project but it's certainly been a learning experience. I've done every aspect of this work in some form previous, entrys, curves, laminations, divided lites, etc. But this one requires putting it all together in a form that was new territory. Luckily I'm not doing this alone, I'm the lead mechanic on this one with a great team of three to four guys. We did a lot of head scratching on round one and went down the occasional dead end technique wise, but it's working now! I appreciate the compliment.

[Joe Calhoon]

Nice work Peter!
You mentioned the stiles would be straight in section. Did this work out or cause any issues?
We are looking at a curved in plan lift - slide job with a tight deadline. Not sure if I am up for the stress.
Joe

[mreza Salav]

Nice work and thanks for sharing it Peter. I'm in the middle of gluing up my entry unit and it has become heavy as hell.

[Peter Quinn]

Thanks for the kind words guys. Mreza, these get pretty heavy with glazing, on this set we are in the nice & light phase, pine core staves and marine fir ply core is all pretty light but strong. I hear you on the weight, yours is a big door to handle solo!

Joe, the flat staves in curved plan proved to be more of a mental problem than a physical one. Once the angles are adjusted you can barely percieve that they are not round, it's a trompe loie of sorts. We made some clamping forms that represent the chord length and height, mostly for clamping and partly for layout. I struck a center line on these and laid the rough parts in order. The rails were cut on the panel saw to length on a curved sled so the blade at 90 degrees cut the rail alone it's radial line as if from the circles center. Put those on the form next to the stiles, shim the stiles so they follow the curve, I needed to adjust the edge of the stile where it meets rail by 1.5 degrees to have its flat segment run parallel to the curve along its section. Me being me wanted to know why.....or develope a consistent way of figuring out the angle precisely rather than manually.

I reasoned that the stiles are 4.5 inches of flat, I came up with a circumference based on the radius, then came up with a ratio of degrees of angle per inch of movement along the circumference. For my radius a move of 4.5 inches along the circumference represented 3 degrees of angle, given 360 in a circle. So I treated it like a miter angle....need to move 3 degrees around a corner, 1.5 degrees on each side. Except the rail angle was established automatically by the sled used to cut it, so I need only put a 1.5 degree angle along the edge of the stile where it met the rail. Then I made a bed board for the sticking on the shaper that lifted the rail to 1.5 degrees so the sticking met the curved rails at the correct angle. Flat stiles appear to follow the curved plan perfectly! Even hanging in the rabbited jambs it's difficult to percieve they are flat, they eye only reads the curved parts, our weatherstrip gap is 3/8" so that hides the subtle deviations.

[peter gagliardi]

Nice work, and nice job photo documenting this. Nobody but the guy/s who make this stuff can understand the time factor involved.

[Peter Quinn]

Here's a little carriage fixture to help get the curved STD (simulated divided lites) bars past the cutter. One wheel has been removed from the feeder, three seemed to distribute pressure unevenly and make the bars lift up, two was just right. The little sub fence in the foreground of the fixture is set friction tight so the small parts can't wander, think of it as a continuous feather board, don't let them move at all. Sanding blocks were made from 3/4" pvc pipe....lots and lots of sanding even though these came out well. Next pic...doors with bars glued in. Sand and straight to finishing....primed and ready to hang. Next stop, build the jambs. Sills were all run on a Hussey as part of round one....does anybody remember how the bottom weather seal goes in? Guess I'll figure that out tomorrow! Good thing for set up blocks.

[Mel Fulks]

Well done and most interesting. Savor every moment!

[mreza Salav]

I always cringe when I see nice wood products get paint over!
Well done sir.

[jack forsberg]

that's real fine work Pete. Thanks for posting this.

[Joe Calhoon]

Thanks for the kind words guys. Mreza, these get pretty heavy with glazing, on this set we are in the nice & light phase, pine core staves and marine fir ply core is all pretty light but strong. I hear you on the weight, yours is a big door to handle solo!

Joe, the flat staves in curved plan proved to be more of a mental problem than a physical one. Once the angles are adjusted you can barely percieve that they are not round, it's a trompe loie of sorts. We made some clamping forms that represent the chord length and height, mostly for clamping and partly for layout. I struck a center line on these and laid the rough parts in order. The rails were cut on the panel saw to length on a curved sled so the blade at 90 degrees cut the rail alone it's radial line as if from the circles center. Put those on the form next to the stiles, shim the stiles so they follow the curve, I needed to adjust the edge of the stile where it meets rail by 1.5 degrees to have its flat segment run parallel to the curve along its section. Me being me wanted to know why.....or develope a consistent way of figuring out the angle precisely rather than manually.

I reasoned that the stiles are 4.5 inches of flat, I came up with a circumference based on the radius, then came up with a ratio of degrees of angle per inch of movement along the circumference. For my radius a move of 4.5 inches along the circumference represented 3 degrees of angle, given 360 in a circle. So I treated it like a miter angle....need to move 3 degrees around a corner, 1.5 degrees on each side. Except the rail angle was established automatically by the sled used to cut it, so I need only put a 1.5 degree angle along the edge of the stile where it met the rail. Then I made a bed board for the sticking on the shaper that lifted the rail to 1.5 degrees so the sticking met the curved rails at the correct angle. Flat stiles appear to follow the curved plan perfectly! Even hanging in the rabbited jambs it's difficult to percieve they are flat, they eye only reads the curved parts, our weatherstrip gap is 3/8" so that hides the subtle deviations.

Thanks Peter, all useful info.

[Peter Quinn]

Thanks for the kind words gentlemen. Joe, it just hit me what you are talking about.....lift and slide.....curve in plan....each door on a slightly different radial line? Ouch, that's a lot of work! Most of these doors I'm making follow the same radius and the pile of jigs and fixtures is bewildering! Now I feel the stress!

[Joe Calhoon]

Thanks for the kind words gentlemen. Joe, it just hit me what you are talking about.....lift and slide.....curve in plan....each door on a slightly different radial line? Ouch, that's a lot of work! Most of these doors I'm making follow the same radius and the pile of jigs and fixtures is bewildering! Now I feel the stress!

Yes, several units all different radius. Big, typ 6'X12' sash. I see the jig work as massive. A dream job if enough time but typical to the construction world schedule is already behind. The radius is big enough that it will be difficult to swing with a trammel. I am going to talk to a CNC guru today to see about making the templates on his machine. Modern miminal design so it won't be quite the details you are dealing with.
Joe

[Peter Quinn]

Its my feeling that in building an entry door the jamb is actually the hardest part to get right. Those transitions from door to framing, weather stripping, water diversion, these are the things that help protect your interior space from the elements to which the door makes you accessible. It helps to have a few good rabbit cutters for this part! And when they say you can never have too many routers, I'm pretty sure they are referring to curved door making. I've tied up almost every router in the shop with some form of fence or guide that really has to remain set...and I've started bringing a few from home. The sills were run on a Hussey molder, they taper 7 degrees or so and curve. Its a two piece threshold with a rabbited bottom, can't put a normal sweep seal in a curved bottom rail, so we went with a double gasket, primary is a silicone sweep, secondary is a silicone flap let into the rabbit in the threshold, the negative rabbit is let into the door bottom, locks it up tight. And how to put all those grooves and rabbits in you ask? Lots and lots of jigs!! A few are pictured here. It took 5 distinct router set ups with special fences or bearings to do the threshold/door bottom prep.

I read the occasional question here about which high priced hinge mortising template to buy...the $299 one from this company or the $400 from that company. Anyone remember the $1.98 beauty contest? Well unlike that show in this case IMO the cheapest solution really is the best, you can easily build one of these precision templates for about $2 worth of MDF and 15 minutes of time. Here's mine, made on a slider or table saw. Note to self....butt template to top of door, remember to add 1/8" shim between dado in jamb legs and template. Always route hinges from the top down, like painting a flag pole. Once hinge mortises are all routed, its time to build the jamb, scribe jamb legs to the threshold, etc. Here's a pair of doors ready to be married. Note the special clamping fixture....flat doors wont slide off the bench while you build the doors up on stickers....but curves will...DAMHIK. :eek:

[Peter Quinn]

that's real fine work Pete. Thanks for posting this.


That means a great deal to me coming from you Jack, thanks for the kind words. Pete

[Keith Mathewson]

Peter I sent you a pm.

Keith

[Peter Quinn]

If the jamb set is the hardest part of the door, then the jamb leg to threshold connection may be the hardest part of the jamb. For water shed you really have to land on top of the threshold, which has to pitch 5-10 degrees down and out depending on who you ask. And this one is a compound angle, curved and beveled. My solution to this tricky intersection is to make a template, or in this case two templates (one right, one left) that fits precisely, then route the jamb legs to fit like a tight scribe. Its hard to get all the nooks and corners with a router, invariably requires some paring and squaring with a chisel. Some tenon the legs to the threshold, here I'm using 3 1/2" #10 GRK exterior screws, very strong, tenacious hold, never see them fail. My strategy to wrap the jamb around the door is to hinge the jamb, hang it, measure the head, install it, measure the threshold, install that with all appropriate gaps for weather seal and clearance. These doors are all pretty precise, so once I have the stops set on the slider it all goes pretty quick.....except the heads and sills are curved, so you need another special fixture and a few tricks to cut the ends of the head/sill parallel to the centerline of the door. On a curved door the jambs don't have to be parallel, they could be splayed to match the walls or follow the radial lines, these are parallel as specified. An extension jamb and trim will splay out to reach the walls and exterior. Lots and lots of trim! Here's a few pics...doors finally hanging..lots of ground left to cover but its a good feeling to reach this part of the job.

[peter gagliardi]

It is interesting that with the in swing of the convex shape that you are able to use a meeting joint that, although stepped is still perpendicular to the curve, and it works! Typically that joint has to be skewed several degrees off perpendicular so they don't interfere.
Again, nice work, it's a shame that it is painted.

[Peter Quinn]

It is interesting that with the in swing of the convex shape that you are able to use a meeting joint that, although stepped is still perpendicular to the curve, and it works! Typically that joint has to be skewed several degrees off perpendicular so they don't interfere.
Again, nice work, it's a shame that it is painted.

Peter, the meeting joint is skewed 11 degrees relative to the center line of the segment to allow the active door to clear the fixed. Our designer animated it in cad and kept skewing it until it worked. Doors are 2 5/16" thick, so a 3 degree bevel wasn't going to work! There will be an applied astragal on the exterior of the inactive side, the offset is around 3/8" plus the gap, there is a double weather gasket in the jog. The bead looks centered, and the astragal will too, but its mostly a trick.

There have been a few comments regarding the fact these are getting painted. I do like the look of a stain grade hardwood door, but these are not going to get the kind of physical protection that clear coated exterior doors require, like a serious portico or porch overhang. So for this application I'm thrilled they are being painted. Like the white paste on a lifeguards nose, a good paint job will give this mahogany a long lifespan with far less maintenance long term IME.

[peter gagliardi]

Ah, I see! Photo makes it look square , that makes more sense now. Thanks.

[Joe Calhoon]

Peter,
I went back through this again. It looks like you made the rails from bender ply with solid skins and edges? Is this correct? Seems like that combined with epoxy or Unibond would make for no springback.
I agree about paint being better in a lot of cases. What usually happens is the client sees the paint grade mahogany and says we cannot paint that. We just started using Accoya for paint grade. Don't have a opinion about it yet.
Joe

[Kevin Jenness]

Beautiful work, Peter. I am curious about the decision to have a recess at the interior of the sill- it seems water exclusion depends on the primary sweep in the door bottom working flawlessly. Are there weep holes in the sill rabbet to allow any water getting in to escape? How do you deal with sealing the area between the pair? Our sills are built with the inner part under the door higher than the beveled outside section, giving a physical barrier to wind driven rain apart from the sweep.

[Mel Fulks]

On the nicer doors I like to sell them on a drip moulding with a fairly sharp edge .

[jack forsberg]

Beautiful work, Peter. I am curious about the decision to have a recess at the interior of the sill- it seems water exclusion depends on the primary sweep in the door bottom working flawlessly. Are there weep holes in the sill rabbet to allow any water getting in to escape? How do you deal with sealing the area between the pair? Our sills are built with the inner part under the door higher than the beveled outside section, giving a physical barrier to wind driven rain apart from the sweep.

I must say that is my thinking too. I do do the rabbited seal but still add a removable adjustable threshold.


When I make the threshold/water sill part of the sill I like it to do a few things for the door. One make it replaceable down the road for ware . Two have it adjustable so I can adjust its thickness so the weatherstripping at the bottom of the door works at its best:p. Three have it out when I hang the door so I don't damage the sill weather striping on the bottom of the door.

we cut the jams joint this way so the sill portion is removable
http://i927.photobucket.com/albums/ad111/tool613/wadkin/doorsill001_zps1af4f2cd.jpg

this lets the threshold sit between the jams for removing/adjusting it or replacement in later years. It also locks the jams in place from moving in on the door. We use brass machine screws taped into the wood to attach this part of the door sill so it can be replaced by the home owner many years latter.i put oil in the threaded hole too. We cover the bottom of our wood sill with ice shield (grace) too so we can level the sill in a bed of caulking and keep moisture from coming up through the bottom from rim linkage condensation or a concrete foundation.

http://i927.photobucket.com/albums/ad111/tool613/wadkin/doorsill004_zps1142d8a8.jpg


I really like to keep my joints in the sill water tight to so I always make provision for large wood movement from wet weather and never have vertical seams in the sill work so water can't sit and rote.

http://i927.photobucket.com/albums/ad111/tool613/wadkin/doorsill002_zpse9deea15.jpg

lastly I put a caulking channels to back up the threshold joint inside/under the threshold out of the sun for a clean water tight threshold joint. This is easy to do by over cutting the threshold rabbit. this does not make it hard to remove or replace it ether but keeps the sill dry from sitting in water that may work its way under it in wet weather that is wind driven other wise.
http://i927.photobucket.com/albums/ad111/tool613/wadkin/doorsill003_zps65edecf6.jpg

there other advantage is it keeps the main sill thick as its not one piece so 8/4 is all that is needed even for wide sells. I do like Mil's idea to of the traditional drip moulding as well.

Very nice work Pete thanks for sharing,